[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M780JR8DFKZ601XAKGTH8ME4":3,"skill-same:01M780JR8DFKZ601XAKGTH8ME4":45},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":24,"commitSha":25,"skillDirectory":26,"forksCount":27,"repositoryHtmlUrl":28,"repositoryHomepageUrl":6,"ownerHtmlUrl":29,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":30,"install":31,"resources":44},"01M780JR8DFKZ601XAKGTH8ME4","effect-v4-source-lookup",null,"Use when you need to confirm an Effect v4 API before relying on it — does this symbol exist, what is its signature, what does it actually do at runtime. Gives the evidence ladder (migration notes settle renames, vendored source settles existence and signature, only a probe settles semantics) and the probe preconditions that keep a probe from silently false-passing against Effect v3.","spencerbeggs\u002Feffected","spencerbeggs","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F109114?v=4","01M77ZTHB70K5Y2VVTRCFY50A5","C. Spencer Beggs","01M77ZTHCC41NXBT3X906K6QM1",8,"MIT","https:\u002F\u002Fgithub.com\u002Fspencerbeggs\u002Feffected\u002Fblob\u002Fmain\u002Fplugin\u002Fskills\u002Feffect-v4-source-lookup\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: effect-v4-source-lookup\ndescription: Use when you need to confirm an Effect v4 API before relying on it — does this symbol exist, what is its signature, what does it actually do at runtime. Gives the evidence ladder (migration notes settle renames, vendored source settles existence and signature, only a probe settles semantics) and the probe preconditions that keep a probe from silently false-passing against Effect v3.\n---\n\n# Looking up the truth about Effect v4\n\nNever write, review, or rely on a v4 API you have not confirmed. Memory is confidently wrong, and so is any `effect@3` a workspace root happens to resolve.\n\nThis skill tells you where to look and how far to go. (If the question is\nmerely \"which module does X\" — start at `effect-v4-module-index`, then verify\nhere.)\n\n**Every claim this skill settles is written about current behaviour, never a\nversion.** Cite a source finding by module and symbol, with the line resolved\nagainst the vendored tree, and never name a pinned Effect version as history —\nthe reader takes their own Effect through their own pin, and a number in a claim\ngoes stale on the next advance without teaching anything the current tree does\nnot already show.\n\n## The evidence ladder\n\nThree rungs, ordered by cost. Each settles a strictly different class of question. Climb to the rung your claim needs — no further, no less.\n\n| Rung | Where | Settles |\n| --- | --- | --- |\n| 1 | `$SRC\u002Fmigration\u002F*.md`, `$SRC\u002Fai-docs\u002F`, `$SRC\u002FLLMS.md` | **Renames** |\n| 2 | `$SRC\u002Fpackages\u002F*\u002Fsrc`, or `$EFFECT_SRC` | **Existence and signature** |\n| 3 | A probe compiled and run from inside a package | **Semantics** |\n\n**Rung 2 also tells you how far to trust an API.** Every module and directly importable export carries an explicit TSDoc `@stability` tag: `stable` follows strict semver, `unstable` may break in a minor release — and the unstable set includes the platform contracts `FileSystem`, `Path`, `PlatformError`, `Stdio` and `Terminal` (`effect-v4-module-index` lists the rest). Grep the source for the tag (`grep -n \"@stability\" $SRC\u002Fpackages\u002Feffect\u002Fsrc\u002F\u003CName>.ts`) before building on an API, and treat an unstable one as a moving target: pin your usage behind one seam and re-verify it when the installed `effect` moves.\n\n### Resolving `$SRC` and `$EFFECT_SRC`\n\nTwo source roots, and they do not settle the same rungs. **Always use the absolute\nform** — the probe protocol below has you `cd` into a package, and every relative\npath breaks the moment you do.\n\n- **`$SRC`** — a vendored Effect checkout, a `git submodule` under `.repos\u002F`\n  pinned to the exact `effect` release tag the workspace compiles against. Serves\n  rungs **1 and 2**. Present when this plugin runs from the `effected` monorepo\n  (after `savvy repos sync` on a fresh clone or worktree); absent elsewhere.\n- **`$EFFECT_SRC`** — `node_modules\u002Feffect\u002Fsrc`. Serves rung **2 only**, and does so\n  at full fidelity: `effect` publishes `src\u002F**\u002F*.ts` in its `files` array, so every\n  consumer already has the complete v4 TypeScript source, `internal\u002F` implementations\n  included. The lockfile pins it to exactly the version you compile against — so on\n  the existence-and-signature question it is at least as trustworthy as the vendored\n  tree, which can drift from the catalog between re-pins. Every `@effect\u002F*` package\n  ships the same way (`@effect\u002Fplatform-node`, `@effect\u002Fvitest`, …), so resolve\n  whichever one your claim is about — the vendored tree's `packages\u002F*\u002Fsrc` has a\n  `node_modules` counterpart in each case. **The converse is false**: the vendored\n  checkout is sparse and carries only `packages\u002Feffect` and `packages\u002Fvitest`, so\n  the platform *implementations* are not there. The Node `FileSystem` — the one\n  `@effected\u002Fmemfs` has to mirror, e.g. `seek` failing `BadArgument` before the\n  start of the file — lives in `@effect\u002Fplatform-node-shared`; under pnpm read\n  `node_modules\u002F.pnpm\u002F@effect+platform-node-shared@\u003Cpin>_effect@\u003Cpin>\u002Fnode_modules\u002F@effect\u002Fplatform-node-shared\u002Fsrc\u002FNodeFileSystem.ts`\n  (the store directory carries the peer suffix; `find node_modules\u002F.pnpm -path\n  '*platform-node-shared@\u003Cpin>*' -name NodeFileSystem.ts` locates it). When a\n  semantic has to be *copied* from the platform layer, that file is rung 2 for it;\n  the vendored tree is silent, not authoritative.\n  **At the same version, `$EFFECT_SRC` and `$SRC` still disagree on line\n  numbers**: npm's published `effect\u002Fsrc` is reformatted at publish time. It\n  carries TSDoc the vendored tag does not, and its signatures are reflowed\n  (a multi-line overload in the tag is joined onto one line in the published\n  copy). A declaration therefore sits at a different line in each tree even\n  where its code is identical and both resolve the same release. A\n  `Module.ts:line` anchor in this plugin always names the **vendored tag**\n  (`$SRC`). Search `node_modules` by symbol name instead of by line —\n  `grep -n \"export const layerStdio\" \"$EFFECT_SRC\u002Fai\u002FMcpServer.ts\"`. The\n  plugin's source repository pins every cited line to its symbol text in\n  `plugin\u002F__test__\u002Fhelpers\u002Fskill-anchors.json`; that manifest is the place\n  to read which symbol an anchor means. Grep for the declaration's name,\n  never for a fragment of a wrapped signature, because the reflow joins\n  lines the tag splits.\n\nResolve both before you trust any lookup. The bottom of the ladder is a hard failure,\nnever a fallback to memory — a wrong answer from v3 memory is indistinguishable from\na right one.\n\n#### When they disagree, the installed source wins\n\n**Do not assume the vendored tree matches what you compile against.** The submodule\nsits at the exact tag it was last pinned to, and a lockfile resolution that moves without its\nmatching `savvy repos pin` leaves the two silently disagreeing — exactly this drift\nhappened under the old subtree pattern, a vendored tree sitting one release\nbehind an already-bumped install — and version-checking is still the reader's job, not\nthe tooling's.\n\nSo a rung-2 answer from `$SRC` can be a *stale* answer, delivered with total confidence\nand no error. The rule:\n\n> **`$EFFECT_SRC` is the authority on existence and signature. `$SRC` is a convenience\n> and the only home of rung 1.** When they disagree, `node_modules` wins — it is what\n> your code links against; the vendored tree is what someone pinned last.\n\nTwo agents were saved by this during the store work. Check the drift in one line\nbefore you trust the tree — the versions match, or they do not (the snippet uses the\n`$SRC` shell variable that the resolution block *below* defines — set it first):\n\n```bash\ndiff \u003C(node -p 'require(\"'\"$SRC\"'\u002Fpackages\u002Feffect\u002Fpackage.json\").version') \\\n     \u003C(node -p 'require(\"effect\u002Fpackage.json\").version') \\\n  || echo \"VENDORED TREE IS STALE — settle rung 2 against \\$EFFECT_SRC, not \\$SRC.\"\n```\n\n**The version-string `diff` above cannot see a second, quieter drift**: even\nwhen the two trees report the identical version, npm's publish step adds\nTSDoc the vendored tag never carries and reflows signatures, so a symbol's\nline number can still differ between them. A version match is not a line-number match — check that\nseparately, against a symbol, never a line count alone:\n\n```bash\ndiff \u003C(grep -n \"export const resource\" \"$SRC\u002Fpackages\u002Feffect\u002Fsrc\u002Fai\u002FMcpServer.ts\") \\\n     \u003C(grep -n \"export const resource\" \"$EFFECT_SRC\u002Fai\u002FMcpServer.ts\") \\\n  || echo \"ANCHORS ARE VENDORED-TREE LINES — search installed source by symbol.\"\n```\n\n```bash\n# Rungs 1+2 — the vendored tree, if this project has one.\nSRC=\"${EFFECT_SMOL_SRC:-${CLAUDE_PROJECT_DIR}\u002F.repos\u002Feffect}\"\ntest -d \"$SRC\u002Fpackages\u002Feffect\u002Fsrc\" || SRC=\"\"\n\n# Rung 2 — the installed source. Resolve it, then GATE ON THE VERSION.\nEFFECT_SRC=\"$(node -p 'const p=require(\"node:path\");p.join(p.dirname(require.resolve(\"effect\u002Fpackage.json\")),\"src\")' 2>\u002Fdev\u002Fnull)\"\ntest -d \"$EFFECT_SRC\" || EFFECT_SRC=\"\"\nif [ -n \"$EFFECT_SRC\" ]; then\n  V=\"$(node -p 'require(\"effect\u002Fpackage.json\").version' 2>\u002Fdev\u002Fnull)\"\n  case \"$V\" in\n    4.*) ;;\n    *) echo \"REFUSING \\$EFFECT_SRC: resolved effect@$V, not v4. cd into a package that depends on effect@4.\" >&2\n       EFFECT_SRC=\"\" ;;\n  esac\nfi\n\nif [ -z \"$SRC\" ] && [ -z \"$EFFECT_SRC\" ]; then\n  echo \"FATAL: no Effect v4 source. Set EFFECT_SMOL_SRC, or cd into a package that depends on effect@4.\" >&2\n  exit 1\nfi\n[ -z \"$SRC\" ] && echo \"NOTE: no vendored tree — rung 1 unavailable. Settle renames at rung 2.\" >&2\n```\n\n**The version gate is the load-bearing line, and it must refuse rather than report.**\n`effect@3` also publishes `src\u002F`, so wherever a v3 is installed, `require.resolve`\nfinds `node_modules\u002F.pnpm\u002Feffect@3.x\u002F...\u002Fsrc` — a complete, confident, *wrong* rung-2\nsource. A block that merely *prints* the resolved version reports v3 source in passing,\nand a reader skims past it. Refuse, and the trap cannot spring.\n\n**That does not mean \"the root gives you v3\".** In *this* repo the\nworkspace root resolves **nothing**: a bare `effect` import there dies with\n`ERR_MODULE_NOT_FOUND`. The lockfile carries **two `effect` entries by intent** —\nthe toolchain's own pin (`@savvy-web\u002Ftsdown-plugins`, `rolldown-pnpm-config` and\n`@vitest-agent\u002F*` declare `effect` and their `@effected\u002F*` inputs as regular\ndependencies, so the published kit binds to the toolchain's copy) and the kit's\nown current pin. Neither copy is v3, and neither resolves from the root; which\none a probe links against is decided by where the probe file lives — which is\nthe point of the rule. `okf\u002Fconventions\u002Fone-resolved-effect-copy.md` has the\nfull shape of the two-copies-by-intent arrangement. Which failure you get\ndepends on what a given repo has installed, so the gate must key on the\n*resolved version*, never on a remembered answer for a particular directory.\n\n**Read the LOCKFILE, and read it carefully — neither the pnpm store nor a raw\ngrep count is the check.** `ls node_modules\u002F.pnpm | grep '^effect@'` lists store\ndirectories, which outlive the lockfile: during a catalog advance the previous\nrelease's directory lingers there, orphaned. And a grep of the lockfile itself\ncan show a second version that is not a second copy — while a toolchain\n`overrides` bridge is up, `pnpm-lock.yaml` carries a redirect line\n(`effect@\u003Cold>: \u003Cnew>`) whose *only* hit is the mapping. One\nresolved version, two spellings. (A `packageExtensions` bridge, the other shape,\nreally does keep two copies — read `okf\u002Fconventions\u002Fone-resolved-effect-copy.md` for which shape stands.) What voids a probe is the version it\n**resolves**, printed from inside itself.\n\n**Rung 1 has no fallback.** The npm package ships no `migration\u002F`, no `ai-docs\u002F`, no\n`LLMS.md` — they are not in its `files` array. That degrades honestly rather than\nsilently: rung 1 is prescriptive, not authoritative, and (as below) a removal is never\nsettled by it anyway. When `$SRC` is absent, say so and climb to rung 2.\n\nThe vendored tree is **read-only**. Never point a writing tool at it: the repo's\nmarkdownlint config sets `\"fix\": true`, and running it over the tree silently rewrites\nthe migration notes. A corrupted rung-1 source cannot be detected by reading it — only\nby `git status \"$SRC\"`.\n\n### Rung 1 is prescriptive, not exhaustive\n\nThe migration notes document the *recommended path*, not the surface. They are excellent for renames and silent about everything else. Verified gaps, each of which cost real work:\n\n- `migration\u002Fservices.md` migrates `Context.Tag` → `Context.Service` and **never mentions `Context.Key`** — which is the primitive you need for a type-only key, and whose `out Shape` covariance means a `Context.Key` parameter will *not* give you the compile error you expect.\n- `migration\u002Fforking.md` names `Effect.fork` → `Effect.forkChild`, and never mentions that `Effect.makeSemaphore` is gone and `Semaphore` is a top-level module (`Semaphore.make` \u002F `makeUnsafe`, then `withPermits(1)(effect)`).\n- `migration\u002Fcause.md` gives `Cause.isFailure` → `Cause.hasFails`, and never mentions `Exit.causeOption` → `Exit.getCause`, which returns `Option\u003CCause\u003CE>>`.\n\nSo: **a removal is never settled by rung 1.** If the docs are silent on a symbol, that is not evidence the symbol is fine. Go to rung 2.\n\n### Rung 1 also asserts things source refutes\n\nSilence is the *gentler* failure. The migration notes also make positive claims that the tree contradicts, in both directions — and a confident wrong answer costs more than an absent one. Both of these were found in one audit and both still hold:\n\n- **A method that does not exist.** `migration\u002Fyieldable.md` documents the `Yieldable` trait as `asEffect(): Effect\u003CA, E, R>` and states the runtime calls `.asEffect()` internally. **`asEffect` has zero occurrences in the entire source tree.** A design built on it fails at the first call.\n- **A removal that did not happen.** `migration\u002Ffiberref.md` lists `Differ` as removed alongside `FiberRef` \u002F `FiberRefs` \u002F `FiberRefsPatch`. Those three are genuinely gone; **`Differ` is alive** (`index.ts:189`), and `migration\u002Fv3-to-v4.md` even maps `effect\u002FDiffer` → `effect\u002FDiffer` and documents the surviving interface. The notes contradict themselves.\n\n**Rung 1 settles renames and nothing else.** Not existence, not removal, not trait mechanics — those are rung 2, and behaviour is rung 3. Treat a positive claim in the notes about *what a symbol is or does* exactly as you treat their silence: unsettled until you have read the source.\n\n### `SCHEMA.md` is a version-exact oracle that still loses to source\n\nThe vendored tree ships `packages\u002Feffect\u002FSCHEMA.md` **at the pin** — 7,400 lines of upstream Schema documentation, versioned with the source rather than floating like a website. That makes it far stronger than `migration\u002F`: it describes the surface you actually have, so it is an excellent **diff oracle** for checking a claim quickly, and when it disagrees with a skill it is usually the skill that is wrong.\n\nIt is still a document, and it does not outrank `src`. An audit of the Schema references found roughly thirty disagreements where `SCHEMA.md` was right — and **eight where it was itself wrong**, on `Getter`\u002F`Parser`, the `effect\u002Fdata` import path, `Schema.brand\u003CT>()`, `UnknownFromJsonString`, `cause.failures`, `Array$`, `toJsonSchema` and `ValidDate`. Those corrections went beyond the pinned upstream doc on source authority.\n\nSo: reach for it early because it is cheap and version-exact, and treat a disagreement with it as a strong signal worth chasing — then settle the answer in `packages\u002Feffect\u002Fsrc`. Call it rung 1.5: better than the migration notes, never a substitute for reading the declaration.\n\n### A bare existence-grep false-negatives on built-in names\n\nRung 2 is usually one grep: `grep -n 'export const Foo\\|export function Foo'`.\nThat grep **lies about names that collide with a TS\u002FJS built-in**, and it lies in\nthe dangerous direction — it reports *absent* for a symbol that exists, which\nreads as \"removed in v4\" and sends you off to invent a replacement.\n\nThe mechanism is that a module cannot declare `const Array` beside its own uses\nof the global `Array` type, so core defines the symbol under a private name and\nrenames it in an `export {}` block:\n\n```text\n\u002F\u002F Schema.ts:4696 — the real definition, under a name you did not grep for\nconst ArraySchema: ArrayLambda = Struct_.lambda\u003CArrayLambda>((schema) => …)\n\n\u002F\u002F Schema.ts:4699 — the export, in a block your grep pattern never matches\nexport { \u002F* …tsdoc… *\u002F ArraySchema as Array }   \u002F\u002F the rename lands at Schema.ts:4717\n```\n\n`Schema.Array` is real, and `grep 'export const Array' Schema.ts` returns\nnothing. The confirmed occurrences of this pattern (vendored-tree lines) —\n`Schema.ts:4717`, `Equivalence.ts:637`, `Order.ts:596` — are all\n`Array`, but treat the *class* of names as suspect, not just that one:\n`Array`, `Record`, `Map`, `Set`, `Error`, `Date`, `Number`, `String`, `Object`,\n`Symbol`, `Function`, `Boolean`. (Some of them do grep normally —\n`Schema.Record` is a plain `export function Record` at `Schema.ts:4032`, and\n`Config.Array` a plain `export function Array` at `Config.ts:1165` — which\nis exactly why the inconsistency catches people.)\n\n**When a built-in-colliding name greps as absent, do not conclude it was\nremoved.** Settle it one of these ways instead:\n\n- **Read the module's export list**, `export {}` blocks included — not just\n  `export const` \u002F `export function` lines.\n- **Grep for the use site, not the declaration**: `grep -rn 'Schema\\.Array('`\n  across this kit's `packages\u002F*\u002Fsrc` proves the call compiles today against the\n  installed `effect`.\n- **Ask the runtime**, from inside a package on the v4 catalog:\n  `node --input-type=module -e \"import * as S from 'effect\u002FSchema'; console.log(typeof S.Array)\"`.\n\nThe same caution applies to the aliased-import direction: this kit writes\n`import { Function as Fn } from \"effect\"`, so `grep 'Function.dual'` misses every\nreal call site, which all read `Fn.dual`.\n\n### The ladder is not Effect-specific: third-party `.d.ts` earns the same climb\n\nEvery rule above is written about `effect` because that is where the cost lands\nmost often, but the *reason* — a confident summary of an API nobody read — has\nnothing to do with which package it is. A dependency's shipped `.d.ts` (or `src`,\nwhere it ships one) is a rung-2 source in exactly the same sense, and it settles\nthe same class of question.\n\nThe concrete case: an evidence pack recommended octokit's\n`RestEndpointMethodTypes` for typing a REST surface, on a plausible reading of\nits docs. **Reading octokit's own types overturned it** — the already-typed\n`request` surface in core was the right seam, and the recommendation would have\nbought a large generated type map for nothing. Nobody had read the `.d.ts`; the\nsummary was reasonable and wrong.\n\n> **Rule: before designing against a third-party type surface, read the surface.**\n> Reach for `node_modules\u002F\u003Cpkg>\u002Fdist\u002F**\u002F*.d.ts` (or its `src`) the same way you\n> reach for `$EFFECT_SRC`. \"The docs say\" is rung 1 for someone else's package\n> too — prescriptive, not exhaustive, and silent about removals.\n\n### The registry rung: published artifacts and installed consumers\n\nThe ladder settles claims about a source tree. Two claim classes need a\ndifferent authority — the registry and the installed artifacts — because no\ntree you are standing in can answer them:\n\n- **Closed upstream ≠ released.** An issue's or PR's closed state proves\n  nothing about any published artifact. What actually shipped is settled by\n  `npm view \u003Cpkg>@\u003Cversion>` and by reading the INSTALLED `.d.ts`\u002F`.js` under\n  `node_modules` — never by the tracker's state.\n- **A repo-local grep structurally cannot see downstream consumers.** Before\n  calling an API change \"breaking in-package only\", read the installed\n  artifacts of the known consumers — the `node_modules` of a consuming repo.\n  That check catches a shape change that would land as a runtime defect in an\n  installed consumer, where a repo-local grep confidently reports zero\n  consumers.\n\n### When two reads of one file disagree, settle it against the committed blob\n\nA second-order failure, and it is a *reasoning* failure rather than a lookup one:\nyou read a file early in a session, read it again later, and the two reads do not\nmatch. The temptation is to pick the read that supports the conclusion you have\nalready drawn — and it is always available, because one of them does.\n\nNeither read is evidence. A working tree changes under you: another agent edits\nit, an earlier step in your own session wrote to it, a formatter ran. **Settle it\nagainst the committed blob** — `git show HEAD:path\u002Fto\u002Ffile` — and then, if the\nworking tree differs, treat that as its own finding to explain rather than\nnoise to discard.\n\nThe rule generalizes past files: **when two observations of the same thing\ndisagree, the tie-break must be a third source, never the one that suits the\nconclusion.** Picking is how a stale read gets laundered into a verified fact.\n\n### Worked example: the three rungs disagree\n\n`Context.Key` (`Context.ts:68`):\n\n- **Rung 1** — `migration\u002Fservices.md` never mentions it. Reading harder produces nothing.\n- **A runtime check** says it does not exist: `typeof Context.Key` is `undefined` and `\"Key\" in Context` is `false`, because it is type-only.\n- **Rung 2** — `$SRC\u002Fpackages\u002Feffect\u002Fsrc\u002FContext.ts:68` settles it:\n\n  ```text\n  export interface Key\u003Cout Identifier, out Shape> extends Effect\u003CShape, never, Identifier>\n  ```\n\n  It exists, it is type-only, and `Shape` is **covariant** — so a `Context.Key` parameter accepts a wider shape than declared, and a design that expected a compile error there will not get one.\n\n  `$EFFECT_SRC\u002FContext.ts:68` is the same declaration, and — for this one\n  anchor — the same line: `Context.ts` overall carries far more publish-time\n  TSDoc in the installed copy than the vendored tag, but the extra lines land\n  after this declaration, not before it. Do not generalize that coincidence:\n  confirm any other anchor by symbol, per the drift check above, rather than\n  assuming a line number survives the same way.\n\nThree answers, one truth, and the cheap rungs are the ones that lie.\n\n### Rung 3 is the only one that settles behaviour\n\nExistence and signature do not tell you what a function *does*. Real examples where the signature was innocent and the behaviour was not:\n\n- `Effect.cached` memoizes the first **failure** as well as the first success, so an `Effect.retry` around the memo replays the cached failure and never re-runs the source. Interruption is the opposite: it is never cached, and an abandoned run is restarted by the next caller.\n- `it.effect` **always** installs a virtual `TestClock`, so `Effect.sleep` \u002F `delay` \u002F `timeout` hang silently to the vitest timeout.\n- `Effect.catchCause` swallows interrupts. A probe built on it reports success exactly where real code hangs. When the exposure is a defect, use `catchDefect`.\n\n## Probe protocol\n\nA probe that cannot fail is worse than no probe. Every precondition below exists because it was violated.\n\n**Venue: a repo with a `scratchpad\u002F` workspace sends probes there first.**\n\n> **\"Scratchpad\" names two unrelated places — do not confuse them.** The agent\n> harness hands most sessions a private scratch DIRECTORY for temporary files\n> (an absolute path under `\u002Ftmp` or similar, named in the system prompt). That\n> is **not** the repo's `scratchpad\u002F` workspace, it has no `node_modules`, and a\n> probe written there dies with\n> `ERR_MODULE_NOT_FOUND: Cannot find package 'effect'` — precondition 3's\n> failure, reached by a route that feels like following this rule rather than\n> breaking it. The venue below means a `scratchpad\u002F` DIRECTORY INSIDE THE REPO,\n> resolved relative to the repo root: a probe written to the harness scratchpad\n> fails to resolve `effect`, and the identical file copied into `packages\u002Fnpm\u002F`\n> runs.\n\nSome kit repos (the effected monorepo among them) ship a private `scratchpad\u002F`\nworkspace member with every kit package at `workspace:*` and `effect` at the\ncatalog's resolution. There, a probe is TYPED: write free-form probes as\n`scratchpad\u002Fprobes\u002F\u003Cname>.ts` and run `pnpm scratchpad:probe probes\u002F\u003Cname>.ts`\nfrom the repo root (`pnpm --filter scratchpad probe` under the hood — pnpm's\n`--filter` executes the underlying script with its cwd inside `scratchpad\u002F`,\nso precondition 1's never-from-the-repo-root resolution hazard does not\narise), or write test-shaped probes as\n`scratchpad\u002F__test__\u002F\u003Cname>.test.ts` and run\n`pnpm exec vitest run --project scratchpad --coverage.enabled=false`. Type-level\nprobes run `pnpm scratchpad:check` (`tsc --noEmit` over the scratchpad\nprogram) — tsx and vitest strip types without checking, so only\n`scratchpad:check` proves a does-this-compile question. Both areas are\ngitignored and disposable (`pnpm scratchpad:reset` reseeds them), so the\nplacement and delete-by-absolute-path rules below — which exist to keep\nprobes safe in repos WITHOUT a scratchpad — do not apply there. Every OTHER\nprecondition (print the resolved version, run the control first, exercise a\nnon-first member) applies unchanged in either venue. No `scratchpad\u002F`\nworkspace in the repo? The protocol below is the venue.\n\n**\"Scratchpad first\" is NOT \"scratchpad always\" — the subject decides.** Read\nwhat the probe is *about* before picking:\n\n| the probe's subject | venue |\n| --- | --- |\n| `effect` itself, or a **published\u002Fbuilt** kit package | the scratchpad. Bare `@effected\u002F\u003Cpkg>` imports resolve to that package's `src\u002F` through the workspace link, and deep-importing `dist\u002Fprod\u002Fnpm\u002Fpkg\u002Findex.js` is the artifact mode for the built answer. |\n| the package's **OWN uncommitted `src\u002F`** | **the package root**, per precondition 3 — a probe file at `packages\u002F\u003Cpkg>\u002Fprobe.ts` importing `.\u002Fsrc\u002F…` relatively. |\n\nThe second row is the one agents get wrong: the scratchpad\nsees committed and uncommitted `src\u002F` alike through the link, but it carries\nits own config and tsconfig, so a probe there does not exercise the package's\nown `tsconfig.json`, `effected.toml` or test setup. Use the package root when\nthe answer depends on those; build with `pnpm build --filter @effected\u002F\u003Cpkg>`\nand use artifact mode when you want to know the *built artifact* behaves that\nway.\n\nTwo riders on the package-root form, both learned by leaving mess behind:\n\n- **A `__probe__\u002F` (or any) subdirectory is wrong twice.** It is\n  precondition 3's silent false-pass — the tsconfig `include` is\n  `${configDir}\u002F*.ts` and does not match subdirectories, so the file leaves the\n  compilation program and its control never fires — and a directory is easier\n  to forget than a file. Three `__probe__\u002F` files were left in\n  `packages\u002Fschemastore\u002F` during one review this way.\n- **Delete by absolute path, and `git status --porcelain packages\u002F\u003Cpkg>` before\n  you report.** The scratchpad's disposability does not transfer: the package\n  root is committed ground.\n\n1. **Run from inside the package, never the repo root.** A workspace root that has a v3 installed resolves it and will describe the v3 surface with total confidence; a root that has none — this repo today — fails with `ERR_MODULE_NOT_FOUND` instead. Both are the same rule: only `packages\u002F\u003Cpkg>\u002F` is guaranteed to resolve the pinned v4.\n2. **Print the resolved version inside every probe, and compare it to the installed `effect` and the vendored tag — never to a prerelease channel word.** `catalog:effected` is a range, so it cannot be the exact match: the exact version is the **lockfile's resolution**, the `version` in `node_modules\u002Feffect\u002Fpackage.json` (and the vendored tree's tag, `.repos\u002Feffect`'s `effect@\u003Cversion>`). The probe prints what *it* resolved and that must equal both; a mismatch means the probe linked a different copy or the vendored tree is stale, and what it says about semantics is unreliable until reconciled. Resolving **v3** (`3.x`) voids a probe outright.\n3. **In a repo without a scratchpad workspace: probe files live at the package root** — *inside* `packages\u002F\u003Cpkg>\u002F`, written there, not merely run from there. Two distinct failures, and they bite at different moments:\n   - **Outside the package, it will not even load.** Node resolves bare imports relative to the **script's own path, not the cwd**, walking up from the file for a `node_modules`. A probe parked in a scratch\u002Ftemp directory therefore dies with `ERR_MODULE_NOT_FOUND: Cannot find package 'effect'` no matter how carefully you `cd packages\u002F\u003Cpkg>` first. Write the file into the package; `cd` alone buys you nothing.\n   - **In a *subdirectory* of the package, it silently false-passes.** The tsconfig `include` is `${configDir}\u002F*.ts` and does **not** match subdirectories, so a probe one level down drops out of the compilation program and its control error never fires.\n\n   The safe spelling is `packages\u002F\u003Cpkg>\u002Fprobe.ts` — package root, top level, deleted by absolute path afterwards.\n\n   **Probing a package's OWN engine\u002Finternal modules: use `npx tsx` — still from a file inside the package.** `npx tsx packages\u002F\u003Cpkg>\u002Fprobe.ts` resolves the house `.js`-extension TypeScript imports correctly and runs without a test project. The file placement rule is NOT relaxed: a probe in `\u002Ftmp` dies with `ERR_MODULE_NOT_FOUND` the moment anything in its import graph names `effect` bare (the entry's own imports resolve from `\u002Ftmp`, which has no `node_modules`). The safe universal spelling is one probe file at the package root, run via `npx tsx`, **never named `*.test.ts`**: one scratch test file with a load-time error fails the package's whole run (`✗ test suite failed to load`, exit 1 — see `effect-v4-testing`). Delete the probe by absolute path afterwards, same as rule 6.\n4. **Run the control first.** Write a line you *know* must fail. Watch it fail. Only then write the real assertion. For a **behavioural** probe, \"must fail\" is the wrong control — invert it and prove the probe can *observe the effect at all*. A probe asking \"does a defect roll the transaction back?\" reads success as *zero rows*, and zero rows is also what a broken harness prints; the control that rescues it is a **committing** transaction that must leave its row behind. Ask what a silently-dead probe would print, and make the control the thing that distinguishes it.\n5. **A probe of any multi-value API must exercise a NON-first member.** A probe that constructs with the first literal of a union, the first element of a list, or the first overload succeeds under both the correct reading and a silently-degraded one — it cannot fail, so it settles nothing. The `@effected\u002Fglob` planning probe for `Schema.Literal(\"a\", \"b\", \"c\")` passed precisely because it constructed with `\"a\"`; only a `\"b\"` construction exposed that v4's runtime keeps the first literal and drops the rest.\n6. **Delete the probe by absolute path** when done.\n\nExample (no-scratchpad venue — see precondition 3):\n\n```bash\ncd packages\u002F\u003Cpkg>\nnode -e 'console.log(\"resolved effect:\", require(\"effect\u002Fpackage.json\").version)'\n# write packages\u002F\u003Cpkg>\u002Fprobe.ts (NOT a temp dir — see precondition 3), then either:\npnpm exec tsc   # type-level probe\nnode probe.ts             # behavioural probe; Node runs TS directly\nrm -f \"$PWD\u002Fprobe.ts\"\n```\n\n**Run `tsc` BARE — never pass the probe file as a CLI argument.** Under\nTypeScript 7 a found tsconfig plus CLI file args is a hard error\n(`TS5112: tsconfig.json is present but will not be loaded if files are\nspecified on commandline`). The bare form compiles the\npackage's own program, which includes a root-level probe (precondition 3's\nwhole point). If a file argument is genuinely unavoidable, `--ignoreConfig`\nproceeds — but it abandons the package's tsconfig, so the bare form stays\ncanonical.\n\nA type-level control that works — shown as `text` because it exists to fail\nthe typecheck:\n\n```text\nimport { Effect } from \"effect\";\nconst control = Effect.catchAll; \u002F\u002F a name that does not exist; must fail\n\u002F\u002F probe.ts(2,24): error TS2339: Property 'catchAll' does not exist on type 'typeof Effect'\n```\n\nInside a probe **file**, print the version with an import, not `require` — a `require`\nalongside a top-level `await` makes the module format ambiguous and Node refuses to run\nit (`ERR_AMBIGUOUS_MODULE_SYNTAX`). The `require` one-liner above is fine because `-e`\nhas no `await`.\n\n```ts\nimport pkg from \"effect\u002Fpackage.json\" with { type: \"json\" };\nconsole.log(\"resolved effect:\", pkg.version); \u002F\u002F must equal the installed (lockfile) effect and the vendored tag\n```\n\n## Portability\n\n`${CLAUDE_PROJECT_DIR}` resolves to the **consuming project's** root, not the plugin's\nrepo. Today those coincide, because the plugin is dogfooded only from the `effected`\nmonorepo — so `$SRC` finds the vendored tree. Published into someone else's project, it\nwill not, and the resolver above is what keeps that from mattering.\n\nWhat a consumer without the vendored tree actually loses is **rung 1 alone**. Rung 2 arrives\nintact via `$EFFECT_SRC`, and rung 3 never needed a tree — a probe compiles against the\npackage the consumer already installed. So the ladder degrades by exactly one rung, and\nby its weakest.\n\n**Never trade the loud failure for a quiet fallback.** Each step above resolves to a\n*version-exact source tree* or it does not resolve. There is no step that resolves to\nrecollection. Silent degradation into memory is the precise failure this plugin\nexists to prevent, and it is indistinguishable from success.\n\nThis is the **only skill that resolves the path** — the agents reference this skill,\nnever the directory. The other skills that mention `.repos\u002Feffect` only *cite* it:\n`effect-v4-mcp` and `effect-v4-cli` name it as the root their `Module.ts:line` anchors\nare pinned to, and `effected-packages`' `pnpm-plugin-effected.md` reference names it as\nthe tree to re-pin when the lockfile's `effect` moves. Keep it that way — a new mention\nunder `plugin\u002Fskills` is either an anchor-root citation like those or a resolver that\nbelongs here. The listing is the positive control, and the filtered count is the gate:\n\n```bash\ngrep -rl '\\.repos\u002Feffect' plugin\u002Fskills | sort   # this skill + the three citing files\ntest \"$(grep -rl '\\.repos\u002Feffect' plugin\u002Fskills \\\n  | grep -v -e effect-v4-source-lookup -e effect-v4-mcp -e effect-v4-cli \\\n            -e references\u002Fpnpm-plugin-effected.md | wc -l)\" -eq 0\n```\n","# Looking up the truth about Effect v4\n\nNever write, review, or rely on a v4 API you have not confirmed. Memory is confidently wrong, and so is any `effect@3` a workspace root happens to resolve.\n\nThis skill tells you where to look and how far to go. (If the question is\nmerely \"which module does X\" — start at `effect-v4-module-index`, then verify\nhere.)\n\n**Every claim this skill settles is written about current behaviour, never a\nversion.** Cite a source finding by module and symbol, with the line resolved\nagainst the vendored tree, and never name a pinned Effect version as history —\nthe reader takes their own Effect through their own pin, and a number in a claim\ngoes stale on the next advance without teaching anything the current tree does\nnot already show.\n\n## The evidence ladder\n\nThree rungs, ordered by cost. Each settles a strictly different class of question. Climb to the rung your claim needs — no further, no less.\n\n| Rung | Where | Settles |\n| --- | --- | --- |\n| 1 | `$SRC\u002Fmigration\u002F*.md`, `$SRC\u002Fai-docs\u002F`, `$SRC\u002FLLMS.md` | **Renames** |\n| 2 | `$SRC\u002Fpackages\u002F*\u002Fsrc`, or `$EFFECT_SRC` | **Existence and signature** |\n| 3 | A probe compiled and run from inside a package | **Semantics** |\n\n**Rung 2 also tells you how far to trust an API.** Every module and directly importable export carries an explicit TSDoc `@stability` tag: `stable` follows strict semver, `unstable` may break in a minor release — and the unstable set includes the platform contracts `FileSystem`, `Path`, `PlatformError`, `Stdio` and `Terminal` (`effect-v4-module-index` lists the rest). Grep the source for the tag (`grep -n \"@stability\" $SRC\u002Fpackages\u002Feffect\u002Fsrc\u002F\u003CName>.ts`) before building on an API, and treat an unstable one as a moving target: pin your usage behind one seam and re-verify it when the installed `effect` moves.\n\n### Resolving `$SRC` and `$EFFECT_SRC`\n\nTwo source roots, and they do not settle the same rungs. **Always use the absolute\nform** — the probe protocol below has you `cd` into a package, and every relative\npath breaks the moment you do.\n\n- **`$SRC`** — a vendored Effect checkout, a `git submodule` under `.repos\u002F`\n  pinned to the exact `effect` release tag the workspace compiles against. Serves\n  rungs **1 and 2**. Present when this plugin runs from the `effected` monorepo\n  (after `savvy repos sync` on a fresh clone or worktree); absent elsewhere.\n- **`$EFFECT_SRC`** — `node_modules\u002Feffect\u002Fsrc`. Serves rung **2 only**, and does so\n  at full fidelity: `effect` publishes `src\u002F**\u002F*.ts` in its `files` array, so every\n  consumer already has the complete v4 TypeScript source, `internal\u002F` implementations\n  included. The lockfile pins it to exactly the version you compile against — so on\n  the existence-and-signature question it is at least as trustworthy as the vendored\n  tree, which can drift from the catalog between re-pins. Every `@effect\u002F*` package\n  ships the same way (`@effect\u002Fplatform-node`, `@effect\u002Fvitest`, …), so resolve\n  whichever one your claim is about — the vendored tree's `packages\u002F*\u002Fsrc` has a\n  `node_modules` counterpart in each case. **The converse is false**: the vendored\n  checkout is sparse and carries only `packages\u002Feffect` and `packages\u002Fvitest`, so\n  the platform *implementations* are not there. The Node `FileSystem` — the one\n  `@effected\u002Fmemfs` has to mirror, e.g. `seek` failing `BadArgument` before the\n  start of the file — lives in `@effect\u002Fplatform-node-shared`; under pnpm read\n  `node_modules\u002F.pnpm\u002F@effect+platform-node-shared@\u003Cpin>_effect@\u003Cpin>\u002Fnode_modules\u002F@effect\u002Fplatform-node-shared\u002Fsrc\u002FNodeFileSystem.ts`\n  (the store directory carries the peer suffix; `find node_modules\u002F.pnpm -path\n  '*platform-node-shared@\u003Cpin>*' -name NodeFileSystem.ts` locates it). When a\n  semantic has to be *copied* from the platform layer, that file is rung 2 for it;\n  the vendored tree is silent, not authoritative.\n  **At the same version, `$EFFECT_SRC` and `$SRC` still disagree on line\n  numbers**: npm's published `effect\u002Fsrc` is reformatted at publish time. It\n  carries TSDoc the vendored tag does not, and its signatures are reflowed\n  (a multi-line overload in the tag is joined onto one line in the published\n  copy). A declaration therefore sits at a different line in each tree even\n  where its code is identical and both resolve the same release. A\n  `Module.ts:line` anchor in this plugin always names the **vendored tag**\n  (`$SRC`). Search `node_modules` by symbol name instead of by line —\n  `grep -n \"export const layerStdio\" \"$EFFECT_SRC\u002Fai\u002FMcpServer.ts\"`. The\n  plugin's source repository pins every cited line to its symbol text in\n  `plugin\u002F__test__\u002Fhelpers\u002Fskill-anchors.json`; that manifest is the place\n  to read which symbol an anchor means. Grep for the declaration's name,\n  never for a fragment of a wrapped signature, because the reflow joins\n  lines the tag splits.\n\nResolve both before you trust any lookup. The bottom of the ladder is a hard failure,\nnever a fallback to memory — a wrong answer from v3 memory is indistinguishable from\na right one.\n\n#### When they disagree, the installed source wins\n\n**Do not assume the vendored tree matches what you compile against.** The submodule\nsits at the exact tag it was last pinned to, and a lockfile resolution that moves without its\nmatching `savvy repos pin` leaves the two silently disagreeing — exactly this drift\nhappened under the old subtree pattern, a vendored tree sitting one release\nbehind an already-bumped install — and version-checking is still the reader's job, not\nthe tooling's.\n\nSo a rung-2 answer from `$SRC` can be a *stale* answer, delivered with total confidence\nand no error. The rule:\n\n> **`$EFFECT_SRC` is the authority on existence and signature. `$SRC` is a convenience\n> and the only home of rung 1.** When they disagree, `node_modules` wins — it is what\n> your code links against; the vendored tree is what someone pinned last.\n\nTwo agents were saved by this during the store work. Check the drift in one line\nbefore you trust the tree — the versions match, or they do not (the snippet uses the\n`$SRC` shell variable that the resolution block *below* defines — set it first):\n\n```bash\ndiff \u003C(node -p 'require(\"'\"$SRC\"'\u002Fpackages\u002Feffect\u002Fpackage.json\").version') \\\n     \u003C(node -p 'require(\"effect\u002Fpackage.json\").version') \\\n  || echo \"VENDORED TREE IS STALE — settle rung 2 against \\$EFFECT_SRC, not \\$SRC.\"\n```\n\n**The version-string `diff` above cannot see a second, quieter drift**: even\nwhen the two trees report the identical version, npm's publish step adds\nTSDoc the vendored tag never carries and reflows signatures, so a symbol's\nline number can still differ between them. A version match is not a line-number match — check that\nseparately, against a symbol, never a line count alone:\n\n```bash\ndiff \u003C(grep -n \"export const resource\" \"$SRC\u002Fpackages\u002Feffect\u002Fsrc\u002Fai\u002FMcpServer.ts\") \\\n     \u003C(grep -n \"export const resource\" \"$EFFECT_SRC\u002Fai\u002FMcpServer.ts\") \\\n  || echo \"ANCHORS ARE VENDORED-TREE LINES — search installed source by symbol.\"\n```\n\n```bash\n# Rungs 1+2 — the vendored tree, if this project has one.\nSRC=\"${EFFECT_SMOL_SRC:-${CLAUDE_PROJECT_DIR}\u002F.repos\u002Feffect}\"\ntest -d \"$SRC\u002Fpackages\u002Feffect\u002Fsrc\" || SRC=\"\"\n\n# Rung 2 — the installed source. Resolve it, then GATE ON THE VERSION.\nEFFECT_SRC=\"$(node -p 'const p=require(\"node:path\");p.join(p.dirname(require.resolve(\"effect\u002Fpackage.json\")),\"src\")' 2>\u002Fdev\u002Fnull)\"\ntest -d \"$EFFECT_SRC\" || EFFECT_SRC=\"\"\nif [ -n \"$EFFECT_SRC\" ]; then\n  V=\"$(node -p 'require(\"effect\u002Fpackage.json\").version' 2>\u002Fdev\u002Fnull)\"\n  case \"$V\" in\n    4.*) ;;\n    *) echo \"REFUSING \\$EFFECT_SRC: resolved effect@$V, not v4. cd into a package that depends on effect@4.\" >&2\n       EFFECT_SRC=\"\" ;;\n  esac\nfi\n\nif [ -z \"$SRC\" ] && [ -z \"$EFFECT_SRC\" ]; then\n  echo \"FATAL: no Effect v4 source. Set EFFECT_SMOL_SRC, or cd into a package that depends on effect@4.\" >&2\n  exit 1\nfi\n[ -z \"$SRC\" ] && echo \"NOTE: no vendored tree — rung 1 unavailable. Settle renames at rung 2.\" >&2\n```\n\n**The version gate is the load-bearing line, and it must refuse rather than report.**\n`effect@3` also publishes `src\u002F`, so wherever a v3 is installed, `require.resolve`\nfinds `node_modules\u002F.pnpm\u002Feffect@3.x\u002F...\u002Fsrc` — a complete, confident, *wrong* rung-2\nsource. A block that merely *prints* the resolved version reports v3 source in passing,\nand a reader skims past it. Refuse, and the trap cannot spring.\n\n**That does not mean \"the root gives you v3\".** In *this* repo the\nworkspace root resolves **nothing**: a bare `effect` import there dies with\n`ERR_MODULE_NOT_FOUND`. The lockfile carries **two `effect` entries by intent** —\nthe toolchain's own pin (`@savvy-web\u002Ftsdown-plugins`, `rolldown-pnpm-config` and\n`@vitest-agent\u002F*` declare `effect` and their `@effected\u002F*` inputs as regular\ndependencies, so the published kit binds to the toolchain's copy) and the kit's\nown current pin. Neither copy is v3, and neither resolves from the root; which\none a probe links against is decided by where the probe file lives — which is\nthe point of the rule. `okf\u002Fconventions\u002Fone-resolved-effect-copy.md` has the\nfull shape of the two-copies-by-intent arrangement. Which failure you get\ndepends on what a given repo has installed, so the gate must key on the\n*resolved version*, never on a remembered answer for a particular directory.\n\n**Read the LOCKFILE, and read it carefully — neither the pnpm store nor a raw\ngrep count is the check.** `ls node_modules\u002F.pnpm | grep '^effect@'` lists store\ndirectories, which outlive the lockfile: during a catalog advance the previous\nrelease's directory lingers there, orphaned. And a grep of the lockfile itself\ncan show a second version that is not a second copy — while a toolchain\n`overrides` bridge is up, `pnpm-lock.yaml` carries a redirect line\n(`effect@\u003Cold>: \u003Cnew>`) whose *only* hit is the mapping. One\nresolved version, two spellings. (A `packageExtensions` bridge, the other shape,\nreally does keep two copies — read `okf\u002Fconventions\u002Fone-resolved-effect-copy.md` for which shape stands.) What voids a probe is the version it\n**resolves**, printed from inside itself.\n\n**Rung 1 has no fallback.** The npm package ships no `migration\u002F`, no `ai-docs\u002F`, no\n`LLMS.md` — they are not in its `files` array. That degrades honestly rather than\nsilently: rung 1 is prescriptive, not authoritative, and (as below) a removal is never\nsettled by it anyway. When `$SRC` is absent, say so and climb to rung 2.\n\nThe vendored tree is **read-only**. Never point a writing tool at it: the repo's\nmarkdownlint config sets `\"fix\": true`, and running it over the tree silently rewrites\nthe migration notes. A corrupted rung-1 source cannot be detected by reading it — only\nby `git status \"$SRC\"`.\n\n### Rung 1 is prescriptive, not exhaustive\n\nThe migration notes document the *recommended path*, not the surface. They are excellent for renames and silent about everything else. Verified gaps, each of which cost real work:\n\n- `migration\u002Fservices.md` migrates `Context.Tag` → `Context.Service` and **never mentions `Context.Key`** — which is the primitive you need for a type-only key, and whose `out Shape` covariance means a `Context.Key` parameter will *not* give you the compile error you expect.\n- `migration\u002Fforking.md` names `Effect.fork` → `Effect.forkChild`, and never mentions that `Effect.makeSemaphore` is gone and `Semaphore` is a top-level module (`Semaphore.make` \u002F `makeUnsafe`, then `withPermits(1)(effect)`).\n- `migration\u002Fcause.md` gives `Cause.isFailure` → `Cause.hasFails`, and never mentions `Exit.causeOption` → `Exit.getCause`, which returns `Option\u003CCause\u003CE>>`.\n\nSo: **a removal is never settled by rung 1.** If the docs are silent on a symbol, that is not evidence the symbol is fine. Go to rung 2.\n\n### Rung 1 also asserts things source refutes\n\nSilence is the *gentler* failure. The migration notes also make positive claims that the tree contradicts, in both directions — and a confident wrong answer costs more than an absent one. Both of these were found in one audit and both still hold:\n\n- **A method that does not exist.** `migration\u002Fyieldable.md` documents the `Yieldable` trait as `asEffect(): Effect\u003CA, E, R>` and states the runtime calls `.asEffect()` internally. **`asEffect` has zero occurrences in the entire source tree.** A design built on it fails at the first call.\n- **A removal that did not happen.** `migration\u002Ffiberref.md` lists `Differ` as removed alongside `FiberRef` \u002F `FiberRefs` \u002F `FiberRefsPatch`. Those three are genuinely gone; **`Differ` is alive** (`index.ts:189`), and `migration\u002Fv3-to-v4.md` even maps `effect\u002FDiffer` → `effect\u002FDiffer` and documents the surviving interface. The notes contradict themselves.\n\n**Rung 1 settles renames and nothing else.** Not existence, not removal, not trait mechanics — those are rung 2, and behaviour is rung 3. Treat a positive claim in the notes about *what a symbol is or does* exactly as you treat their silence: unsettled until you have read the source.\n\n### `SCHEMA.md` is a version-exact oracle that still loses to source\n\nThe vendored tree ships `packages\u002Feffect\u002FSCHEMA.md` **at the pin** — 7,400 lines of upstream Schema documentation, versioned with the source rather than floating like a website. That makes it far stronger than `migration\u002F`: it describes the surface you actually have, so it is an excellent **diff oracle** for checking a claim quickly, and when it disagrees with a skill it is usually the skill that is wrong.\n\nIt is still a document, and it does not outrank `src`. An audit of the Schema references found roughly thirty disagreements where `SCHEMA.md` was right — and **eight where it was itself wrong**, on `Getter`\u002F`Parser`, the `effect\u002Fdata` import path, `Schema.brand\u003CT>()`, `UnknownFromJsonString`, `cause.failures`, `Array$`, `toJsonSchema` and `ValidDate`. Those corrections went beyond the pinned upstream doc on source authority.\n\nSo: reach for it early because it is cheap and version-exact, and treat a disagreement with it as a strong signal worth chasing — then settle the answer in `packages\u002Feffect\u002Fsrc`. Call it rung 1.5: better than the migration notes, never a substitute for reading the declaration.\n\n### A bare existence-grep false-negatives on built-in names\n\nRung 2 is usually one grep: `grep -n 'export const Foo\\|export function Foo'`.\nThat grep **lies about names that collide with a TS\u002FJS built-in**, and it lies in\nthe dangerous direction — it reports *absent* for a symbol that exists, which\nreads as \"removed in v4\" and sends you off to invent a replacement.\n\nThe mechanism is that a module cannot declare `const Array` beside its own uses\nof the global `Array` type, so core defines the symbol under a private name and\nrenames it in an `export {}` block:\n\n```text\n\u002F\u002F Schema.ts:4696 — the real definition, under a name you did not grep for\nconst ArraySchema: ArrayLambda = Struct_.lambda\u003CArrayLambda>((schema) => …)\n\n\u002F\u002F Schema.ts:4699 — the export, in a block your grep pattern never matches\nexport { \u002F* …tsdoc… *\u002F ArraySchema as Array }   \u002F\u002F the rename lands at Schema.ts:4717\n```\n\n`Schema.Array` is real, and `grep 'export const Array' Schema.ts` returns\nnothing. The confirmed occurrences of this pattern (vendored-tree lines) —\n`Schema.ts:4717`, `Equivalence.ts:637`, `Order.ts:596` — are all\n`Array`, but treat the *class* of names as suspect, not just that one:\n`Array`, `Record`, `Map`, `Set`, `Error`, `Date`, `Number`, `String`, `Object`,\n`Symbol`, `Function`, `Boolean`. (Some of them do grep normally —\n`Schema.Record` is a plain `export function Record` at `Schema.ts:4032`, and\n`Config.Array` a plain `export function Array` at `Config.ts:1165` — which\nis exactly why the inconsistency catches people.)\n\n**When a built-in-colliding name greps as absent, do not conclude it was\nremoved.** Settle it one of these ways instead:\n\n- **Read the module's export list**, `export {}` blocks included — not just\n  `export const` \u002F `export function` lines.\n- **Grep for the use site, not the declaration**: `grep -rn 'Schema\\.Array('`\n  across this kit's `packages\u002F*\u002Fsrc` proves the call compiles today against the\n  installed `effect`.\n- **Ask the runtime**, from inside a package on the v4 catalog:\n  `node --input-type=module -e \"import * as S from 'effect\u002FSchema'; console.log(typeof S.Array)\"`.\n\nThe same caution applies to the aliased-import direction: this kit writes\n`import { Function as Fn } from \"effect\"`, so `grep 'Function.dual'` misses every\nreal call site, which all read `Fn.dual`.\n\n### The ladder is not Effect-specific: third-party `.d.ts` earns the same climb\n\nEvery rule above is written about `effect` because that is where the cost lands\nmost often, but the *reason* — a confident summary of an API nobody read — has\nnothing to do with which package it is. A dependency's shipped `.d.ts` (or `src`,\nwhere it ships one) is a rung-2 source in exactly the same sense, and it settles\nthe same class of question.\n\nThe concrete case: an evidence pack recommended octokit's\n`RestEndpointMethodTypes` for typing a REST surface, on a plausible reading of\nits docs. **Reading octokit's own types overturned it** — the already-typed\n`request` surface in core was the right seam, and the recommendation would have\nbought a large generated type map for nothing. Nobody had read the `.d.ts`; the\nsummary was reasonable and wrong.\n\n> **Rule: before designing against a third-party type surface, read the surface.**\n> Reach for `node_modules\u002F\u003Cpkg>\u002Fdist\u002F**\u002F*.d.ts` (or its `src`) the same way you\n> reach for `$EFFECT_SRC`. \"The docs say\" is rung 1 for someone else's package\n> too — prescriptive, not exhaustive, and silent about removals.\n\n### The registry rung: published artifacts and installed consumers\n\nThe ladder settles claims about a source tree. Two claim classes need a\ndifferent authority — the registry and the installed artifacts — because no\ntree you are standing in can answer them:\n\n- **Closed upstream ≠ released.** An issue's or PR's closed state proves\n  nothing about any published artifact. What actually shipped is settled by\n  `npm view \u003Cpkg>@\u003Cversion>` and by reading the INSTALLED `.d.ts`\u002F`.js` under\n  `node_modules` — never by the tracker's state.\n- **A repo-local grep structurally cannot see downstream consumers.** Before\n  calling an API change \"breaking in-package only\", read the installed\n  artifacts of the known consumers — the `node_modules` of a consuming repo.\n  That check catches a shape change that would land as a runtime defect in an\n  installed consumer, where a repo-local grep confidently reports zero\n  consumers.\n\n### When two reads of one file disagree, settle it against the committed blob\n\nA second-order failure, and it is a *reasoning* failure rather than a lookup one:\nyou read a file early in a session, read it again later, and the two reads do not\nmatch. The temptation is to pick the read that supports the conclusion you have\nalready drawn — and it is always available, because one of them does.\n\nNeither read is evidence. A working tree changes under you: another agent edits\nit, an earlier step in your own session wrote to it, a formatter ran. **Settle it\nagainst the committed blob** — `git show HEAD:path\u002Fto\u002Ffile` — and then, if the\nworking tree differs, treat that as its own finding to explain rather than\nnoise to discard.\n\nThe rule generalizes past files: **when two observations of the same thing\ndisagree, the tie-break must be a third source, never the one that suits the\nconclusion.** Picking is how a stale read gets laundered into a verified fact.\n\n### Worked example: the three rungs disagree\n\n`Context.Key` (`Context.ts:68`):\n\n- **Rung 1** — `migration\u002Fservices.md` never mentions it. Reading harder produces nothing.\n- **A runtime check** says it does not exist: `typeof Context.Key` is `undefined` and `\"Key\" in Context` is `false`, because it is type-only.\n- **Rung 2** — `$SRC\u002Fpackages\u002Feffect\u002Fsrc\u002FContext.ts:68` settles it:\n\n  ```text\n  export interface Key\u003Cout Identifier, out Shape> extends Effect\u003CShape, never, Identifier>\n  ```\n\n  It exists, it is type-only, and `Shape` is **covariant** — so a `Context.Key` parameter accepts a wider shape than declared, and a design that expected a compile error there will not get one.\n\n  `$EFFECT_SRC\u002FContext.ts:68` is the same declaration, and — for this one\n  anchor — the same line: `Context.ts` overall carries far more publish-time\n  TSDoc in the installed copy than the vendored tag, but the extra lines land\n  after this declaration, not before it. Do not generalize that coincidence:\n  confirm any other anchor by symbol, per the drift check above, rather than\n  assuming a line number survives the same way.\n\nThree answers, one truth, and the cheap rungs are the ones that lie.\n\n### Rung 3 is the only one that settles behaviour\n\nExistence and signature do not tell you what a function *does*. Real examples where the signature was innocent and the behaviour was not:\n\n- `Effect.cached` memoizes the first **failure** as well as the first success, so an `Effect.retry` around the memo replays the cached failure and never re-runs the source. Interruption is the opposite: it is never cached, and an abandoned run is restarted by the next caller.\n- `it.effect` **always** installs a virtual `TestClock`, so `Effect.sleep` \u002F `delay` \u002F `timeout` hang silently to the vitest timeout.\n- `Effect.catchCause` swallows interrupts. A probe built on it reports success exactly where real code hangs. When the exposure is a defect, use `catchDefect`.\n\n## Probe protocol\n\nA probe that cannot fail is worse than no probe. Every precondition below exists because it was violated.\n\n**Venue: a repo with a `scratchpad\u002F` workspace sends probes there first.**\n\n> **\"Scratchpad\" names two unrelated places — do not confuse them.** The agent\n> harness hands most sessions a private scratch DIRECTORY for temporary files\n> (an absolute path under `\u002Ftmp` or similar, named in the system prompt). That\n> is **not** the repo's `scratchpad\u002F` workspace, it has no `node_modules`, and a\n> probe written there dies with\n> `ERR_MODULE_NOT_FOUND: Cannot find package 'effect'` — precondition 3's\n> failure, reached by a route that feels like following this rule rather than\n> breaking it. The venue below means a `scratchpad\u002F` DIRECTORY INSIDE THE REPO,\n> resolved relative to the repo root: a probe written to the harness scratchpad\n> fails to resolve `effect`, and the identical file copied into `packages\u002Fnpm\u002F`\n> runs.\n\nSome kit repos (the effected monorepo among them) ship a private `scratchpad\u002F`\nworkspace member with every kit package at `workspace:*` and `effect` at the\ncatalog's resolution. There, a probe is TYPED: write free-form probes as\n`scratchpad\u002Fprobes\u002F\u003Cname>.ts` and run `pnpm scratchpad:probe probes\u002F\u003Cname>.ts`\nfrom the repo root (`pnpm --filter scratchpad probe` under the hood — pnpm's\n`--filter` executes the underlying script with its cwd inside `scratchpad\u002F`,\nso precondition 1's never-from-the-repo-root resolution hazard does not\narise), or write test-shaped probes as\n`scratchpad\u002F__test__\u002F\u003Cname>.test.ts` and run\n`pnpm exec vitest run --project scratchpad --coverage.enabled=false`. Type-level\nprobes run `pnpm scratchpad:check` (`tsc --noEmit` over the scratchpad\nprogram) — tsx and vitest strip types without checking, so only\n`scratchpad:check` proves a does-this-compile question. Both areas are\ngitignored and disposable (`pnpm scratchpad:reset` reseeds them), so the\nplacement and delete-by-absolute-path rules below — which exist to keep\nprobes safe in repos WITHOUT a scratchpad — do not apply there. Every OTHER\nprecondition (print the resolved version, run the control first, exercise a\nnon-first member) applies unchanged in either venue. No `scratchpad\u002F`\nworkspace in the repo? The protocol below is the venue.\n\n**\"Scratchpad first\" is NOT \"scratchpad always\" — the subject decides.** Read\nwhat the probe is *about* before picking:\n\n| the probe's subject | venue |\n| --- | --- |\n| `effect` itself, or a **published\u002Fbuilt** kit package | the scratchpad. Bare `@effected\u002F\u003Cpkg>` imports resolve to that package's `src\u002F` through the workspace link, and deep-importing `dist\u002Fprod\u002Fnpm\u002Fpkg\u002Findex.js` is the artifact mode for the built answer. |\n| the package's **OWN uncommitted `src\u002F`** | **the package root**, per precondition 3 — a probe file at `packages\u002F\u003Cpkg>\u002Fprobe.ts` importing `.\u002Fsrc\u002F…` relatively. |\n\nThe second row is the one agents get wrong: the scratchpad\nsees committed and uncommitted `src\u002F` alike through the link, but it carries\nits own config and tsconfig, so a probe there does not exercise the package's\nown `tsconfig.json`, `effected.toml` or test setup. Use the package root when\nthe answer depends on those; build with `pnpm build --filter @effected\u002F\u003Cpkg>`\nand use artifact mode when you want to know the *built artifact* behaves that\nway.\n\nTwo riders on the package-root form, both learned by leaving mess behind:\n\n- **A `__probe__\u002F` (or any) subdirectory is wrong twice.** It is\n  precondition 3's silent false-pass — the tsconfig `include` is\n  `${configDir}\u002F*.ts` and does not match subdirectories, so the file leaves the\n  compilation program and its control never fires — and a directory is easier\n  to forget than a file. Three `__probe__\u002F` files were left in\n  `packages\u002Fschemastore\u002F` during one review this way.\n- **Delete by absolute path, and `git status --porcelain packages\u002F\u003Cpkg>` before\n  you report.** The scratchpad's disposability does not transfer: the package\n  root is committed ground.\n\n1. **Run from inside the package, never the repo root.** A workspace root that has a v3 installed resolves it and will describe the v3 surface with total confidence; a root that has none — this repo today — fails with `ERR_MODULE_NOT_FOUND` instead. Both are the same rule: only `packages\u002F\u003Cpkg>\u002F` is guaranteed to resolve the pinned v4.\n2. **Print the resolved version inside every probe, and compare it to the installed `effect` and the vendored tag — never to a prerelease channel word.** `catalog:effected` is a range, so it cannot be the exact match: the exact version is the **lockfile's resolution**, the `version` in `node_modules\u002Feffect\u002Fpackage.json` (and the vendored tree's tag, `.repos\u002Feffect`'s `effect@\u003Cversion>`). The probe prints what *it* resolved and that must equal both; a mismatch means the probe linked a different copy or the vendored tree is stale, and what it says about semantics is unreliable until reconciled. Resolving **v3** (`3.x`) voids a probe outright.\n3. **In a repo without a scratchpad workspace: probe files live at the package root** — *inside* `packages\u002F\u003Cpkg>\u002F`, written there, not merely run from there. Two distinct failures, and they bite at different moments:\n   - **Outside the package, it will not even load.** Node resolves bare imports relative to the **script's own path, not the cwd**, walking up from the file for a `node_modules`. A probe parked in a scratch\u002Ftemp directory therefore dies with `ERR_MODULE_NOT_FOUND: Cannot find package 'effect'` no matter how carefully you `cd packages\u002F\u003Cpkg>` first. Write the file into the package; `cd` alone buys you nothing.\n   - **In a *subdirectory* of the package, it silently false-passes.** The tsconfig `include` is `${configDir}\u002F*.ts` and does **not** match subdirectories, so a probe one level down drops out of the compilation program and its control error never fires.\n\n   The safe spelling is `packages\u002F\u003Cpkg>\u002Fprobe.ts` — package root, top level, deleted by absolute path afterwards.\n\n   **Probing a package's OWN engine\u002Finternal modules: use `npx tsx` — still from a file inside the package.** `npx tsx packages\u002F\u003Cpkg>\u002Fprobe.ts` resolves the house `.js`-extension TypeScript imports correctly and runs without a test project. The file placement rule is NOT relaxed: a probe in `\u002Ftmp` dies with `ERR_MODULE_NOT_FOUND` the moment anything in its import graph names `effect` bare (the entry's own imports resolve from `\u002Ftmp`, which has no `node_modules`). The safe universal spelling is one probe file at the package root, run via `npx tsx`, **never named `*.test.ts`**: one scratch test file with a load-time error fails the package's whole run (`✗ test suite failed to load`, exit 1 — see `effect-v4-testing`). Delete the probe by absolute path afterwards, same as rule 6.\n4. **Run the control first.** Write a line you *know* must fail. Watch it fail. Only then write the real assertion. For a **behavioural** probe, \"must fail\" is the wrong control — invert it and prove the probe can *observe the effect at all*. A probe asking \"does a defect roll the transaction back?\" reads success as *zero rows*, and zero rows is also what a broken harness prints; the control that rescues it is a **committing** transaction that must leave its row behind. Ask what a silently-dead probe would print, and make the control the thing that distinguishes it.\n5. **A probe of any multi-value API must exercise a NON-first member.** A probe that constructs with the first literal of a union, the first element of a list, or the first overload succeeds under both the correct reading and a silently-degraded one — it cannot fail, so it settles nothing. The `@effected\u002Fglob` planning probe for `Schema.Literal(\"a\", \"b\", \"c\")` passed precisely because it constructed with `\"a\"`; only a `\"b\"` construction exposed that v4's runtime keeps the first literal and drops the rest.\n6. **Delete the probe by absolute path** when done.\n\nExample (no-scratchpad venue — see precondition 3):\n\n```bash\ncd packages\u002F\u003Cpkg>\nnode -e 'console.log(\"resolved effect:\", require(\"effect\u002Fpackage.json\").version)'\n# write packages\u002F\u003Cpkg>\u002Fprobe.ts (NOT a temp dir — see precondition 3), then either:\npnpm exec tsc   # type-level probe\nnode probe.ts             # behavioural probe; Node runs TS directly\nrm -f \"$PWD\u002Fprobe.ts\"\n```\n\n**Run `tsc` BARE — never pass the probe file as a CLI argument.** Under\nTypeScript 7 a found tsconfig plus CLI file args is a hard error\n(`TS5112: tsconfig.json is present but will not be loaded if files are\nspecified on commandline`). The bare form compiles the\npackage's own program, which includes a root-level probe (precondition 3's\nwhole point). If a file argument is genuinely unavoidable, `--ignoreConfig`\nproceeds — but it abandons the package's tsconfig, so the bare form stays\ncanonical.\n\nA type-level control that works — shown as `text` because it exists to fail\nthe typecheck:\n\n```text\nimport { Effect } from \"effect\";\nconst control = Effect.catchAll; \u002F\u002F a name that does not exist; must fail\n\u002F\u002F probe.ts(2,24): error TS2339: Property 'catchAll' does not exist on type 'typeof Effect'\n```\n\nInside a probe **file**, print the version with an import, not `require` — a `require`\nalongside a top-level `await` makes the module format ambiguous and Node refuses to run\nit (`ERR_AMBIGUOUS_MODULE_SYNTAX`). The `require` one-liner above is fine because `-e`\nhas no `await`.\n\n```ts\nimport pkg from \"effect\u002Fpackage.json\" with { type: \"json\" };\nconsole.log(\"resolved effect:\", pkg.version); \u002F\u002F must equal the installed (lockfile) effect and the vendored tag\n```\n\n## Portability\n\n`${CLAUDE_PROJECT_DIR}` resolves to the **consuming project's** root, not the plugin's\nrepo. Today those coincide, because the plugin is dogfooded only from the `effected`\nmonorepo — so `$SRC` finds the vendored tree. Published into someone else's project, it\nwill not, and the resolver above is what keeps that from mattering.\n\nWhat a consumer without the vendored tree actually loses is **rung 1 alone**. Rung 2 arrives\nintact via `$EFFECT_SRC`, and rung 3 never needed a tree — a probe compiles against the\npackage the consumer already installed. So the ladder degrades by exactly one rung, and\nby its weakest.\n\n**Never trade the loud failure for a quiet fallback.** Each step above resolves to a\n*version-exact source tree* or it does not resolve. There is no step that resolves to\nrecollection. Silent degradation into memory is the precise failure this plugin\nexists to prevent, and it is indistinguishable from success.\n\nThis is the **only skill that resolves the path** — the agents reference this skill,\nnever the directory. The other skills that mention `.repos\u002Feffect` only *cite* it:\n`effect-v4-mcp` and `effect-v4-cli` name it as the root their `Module.ts:line` anchors\nare pinned to, and `effected-packages`' `pnpm-plugin-effected.md` reference names it as\nthe tree to re-pin when the lockfile's `effect` moves. Keep it that way — a new mention\nunder `plugin\u002Fskills` is either an anchor-root citation like those or a resolver that\nbelongs here. The listing is the positive control, and the filtered count is the gate:\n\n```bash\ngrep -rl '\\.repos\u002Feffect' plugin\u002Fskills | sort   # this skill + the three citing files\ntest \"$(grep -rl '\\.repos\u002Feffect' plugin\u002Fskills \\\n  | grep -v -e effect-v4-source-lookup -e effect-v4-mcp -e effect-v4-cli \\\n            -e references\u002Fpnpm-plugin-effected.md | wc -l)\" -eq 0\n```\n",{"name":5,"description":7},"main","11d6b3ae59538b0e16ad67c315ecb0c25fcb0125","plugin\u002Fskills\u002Feffect-v4-source-lookup",2,"https:\u002F\u002Fgithub.com\u002Fspencerbeggs\u002Feffected","https:\u002F\u002Fgithub.com\u002Fspencerbeggs","2026-11-14 07:03:25",{"source":8,"skillName":5,"commands":32,"prompt":42,"zipAvailable":43},[33,36,39],{"tool":34,"command":35},"npx","npx skills add spencerbeggs\u002Feffected --skill effect-v4-source-lookup",{"tool":37,"command":38},"bunx","bunx skills add spencerbeggs\u002Feffected --skill effect-v4-source-lookup",{"tool":40,"command":41},"pnpm","pnpm dlx skills add spencerbeggs\u002Feffected --skill effect-v4-source-lookup","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fspencerbeggs\u002Feffected\u002Fblob\u002Fmain\u002Fplugin\u002Fskills\u002Feffect-v4-source-lookup\u002FSKILL.md",true,[],[46,53,60,67,74,81],{"skillNo":47,"name":48,"nameZh":6,"description":49,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":50,"categories":51,"keywords":52,"publishedAt":19},"01M780NVTD800KMFV869YNZED6","testing-actions","Use when writing or reviewing tests for @effected\u002Fgithub-actions, @effected\u002Fgithub, @effected\u002Fsbom or @effected\u002Fcommands — stubbing the GitHub API, writing a service's layerTest or makeTest double, testing the pre\u002Fmain\u002Fpost lifecycle through Action.run, or debugging why a test hits the real network or hangs.","https:\u002F\u002Fgithub.com\u002Fspencerbeggs\u002Feffected\u002Fblob\u002Fmain\u002Fplugin\u002Fskills\u002Ftesting-actions\u002FSKILL.md",[],[],{"skillNo":54,"name":55,"nameZh":6,"description":56,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":57,"categories":58,"keywords":59,"publishedAt":19},"01M780NPDSGFC0FCP7696NW11J","supply-chain-attestation","Use when generating a CycloneDX SBOM, checking NTIA minimum elements, building an in-toto statement or a SLSA provenance predicate, signing into a Sigstore DSSE bundle, or uploading an attestation to GitHub.","https:\u002F\u002Fgithub.com\u002Fspencerbeggs\u002Feffected\u002Fblob\u002Fmain\u002Fplugin\u002Fskills\u002Fsupply-chain-attestation\u002FSKILL.md",[],[],{"skillNo":61,"name":62,"nameZh":6,"description":63,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":64,"categories":65,"keywords":66,"publishedAt":19},"01M780NGP4QNZDH8T8Q8WZ18SA","structuring-an-action","Use when laying out a new GitHub Action repository on the @effected suite, deciding where a piece of code belongs (an entry point, a step, a shared service, a shim), or reviewing an existing action's file structure against the canonical shape.","https:\u002F\u002Fgithub.com\u002Fspencerbeggs\u002Feffected\u002Fblob\u002Fmain\u002Fplugin\u002Fskills\u002Fstructuring-an-action\u002FSKILL.md",[],[],{"skillNo":68,"name":69,"nameZh":6,"description":70,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":71,"categories":72,"keywords":73,"publishedAt":19},"01M780NCD52H1Y03FS83WJ8BT5","running-commands-and-tools","Use when running a subprocess in Effect v4, spawning a command, capturing stdout\u002Fstderr\u002Fexit code, checking whether a CLI tool is installed or which copy to run, running a package-manager script through the project's launcher, detaching a background process, or redacting a secret from argv or captured output.","https:\u002F\u002Fgithub.com\u002Fspencerbeggs\u002Feffected\u002Fblob\u002Fmain\u002Fplugin\u002Fskills\u002Frunning-commands-and-tools\u002FSKILL.md",[],[],{"skillNo":75,"name":76,"nameZh":6,"description":77,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":78,"categories":79,"keywords":80,"publishedAt":19},"01M780N823DWNRMSY5X2NT5MSR","release-and-publish","Use when publishing a package to npm, reading the npm registry from Effect v4, checking whether a version is already published, cutting a release tag or GitHub release from an action, applying a release-age gate, or deriving tracking tags for GitHub Actions distribution.","https:\u002F\u002Fgithub.com\u002Fspencerbeggs\u002Feffected\u002Fblob\u002Fmain\u002Fplugin\u002Fskills\u002Frelease-and-publish\u002FSKILL.md",[],[],{"skillNo":82,"name":83,"nameZh":6,"description":84,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":85,"categories":86,"keywords":87,"publishedAt":19},"01M780MXRF6H20KCESTSSFB3HQ","hardening-a-parser-port","Use when porting or writing a recursive-descent parser, lexer, or tree-walker over untrusted text in the @effected monorepo — the class of hardening the cloud reviewer scans for on every migration. Covers stack-overflow depth guards (on EVERY recursion surface, which a facade has N of — not two), numeric bound guards that must reject NaN and non-integers, code-point range checks scoped to formats with wide escapes, prototype-pollution, control-character rejection, and the invariant that malformed input must fail through the typed error channel, never as an unhandled defect.","https:\u002F\u002Fgithub.com\u002Fspencerbeggs\u002Feffected\u002Fblob\u002Fmain\u002Fplugin\u002Fskills\u002Fhardening-a-parser-port\u002FSKILL.md",[],[]]